Polypharmacy was first described in the psychiatric literature in 1969. ^{Reference Sheppard, Collins, Fiorentino, Fracchia and Merlis1} Studies reported variable rates of concurrent antipsychotic prescription depending on the population considered. A study in Australia, examining people receiving out-patient treatment for schizophrenia, showed a 13% rate of multiple antipsychotic prescription use. ^{Reference Keks, Alston, Hope, Krapivensky, Culhane and Tanaghow2} One Japanese study indicated that the rate of antipsychotic polypharmacy exceeded 90%. ^{Reference Ito, Kubota and Sato3} A recent study in the UK showed an intermediate rate of 30%. ^{Reference Taylor, Shajahan and Lawrie4} Despite its common occurrence, the evidence base behind antipsychotic polypharmacy is widely recognised as limited and its use has been considered both a ‘therapeutic option’ and a ‘dirty little secret’. ^{Reference Stahl5}

Routes to polypharmacy

When considering the prescription of more than one antipsychotic, perceived potential clinical benefit rather than receptor theory often leads to the chosen combination. It is increasingly evident that various non-dopaminergic receptors have an important role in the clinical profile of schizophrenia –with adrenergic, glutamatergic and serotonergic receptors (in particular 5-HT_2A receptors) involved in the pathogenesis of positive and negative symptoms. It is postulated that combined therapy might help to diversify the range of receptors affected by drug treatment and thereby improve symptoms of disease and reduce recurrence rates. ^{Reference Freudenreich and Goff6}

The National Institute for Health and Clinical Excellence (NICE) has produced algorithms for the use of antipsychotics in schizophrenia. ⁷ The guidance states that in a first presentation, concurrent antipsychotics should not be prescribed except to cover short change-over periods (i.e. cross-tapering) and recommends that a second antipsychotic could be ‘cautiously’ trialled in combination with clozapine in patients whose disorder is truly ‘treatment resistant’, following an inadequate response to adequate trials of initial antipsychotics and also clozapine. It is only in these specific clinical settings that NICE suggests dual antipsychotic prescription should occur.

However, antipsychotic polypharmacy prescription may occur in other clinical scenarios. When clinicians aim to switch their patients from one antipsychotic to another by cross-tapering, the switch may never be fully completed as symptoms stabilise or improve. This leads to reluctance on the part of clinician or patient to complete the switch and may postpone the completion of the switch indefinitely, leading to long-term antipsychotic polypharmacy.

Clinicians sometimes prescribe antipsychotic doses close to or beyond that defined as maximum by the UK British National Formulary (BNF). ⁸ When this occurs, high-dose monitoring protocols are recommended. ^{Reference Taylor, Paton and Kerwin9} Good clinical practice suggests that we avoid high-dose prescribing in order to minimise side-effects. For this reason, clinicians may trial more than one antipsychotic at lower doses, leading once again to long-term antipsychotic polypharmacy. However, polypharmacy may in itself unwittingly result in high total dose prescribing. An audit carried out by the Prescribing Observatory for Mental Health found that 80% of patients prescribed combined antipsychotics, including those prescribed them on an ‘as required’ basis, had a total dose above the recommended limits by the BNF. ¹⁰ It is unclear if this incidence of high total dose prescribing within polypharmacy use is representative of current practice.

Another possible route to antipsychotic polypharmacy is that ‘stable’ patients with psychosis, receiving long-term antipsychotic monotherapy, may develop an acute exacerbation of psychotic symptoms, or worsening of agitation or insomnia and be prescribed an additional antipsychotic on an ‘as required’ or short-term basis. However, such short-term or ‘as required’ prescriptions may develop into medium- or long-term adjuvant treatments, particularly if considered successful in the short term. A natural reluctance to risk a recurrence of symptoms tends to perpetuate this situation and ultimately results in long-term antipsychotic polypharmacy.

Evidence of efficacy

There are only a small number of randomised controlled trials comparing monotherapy with polypharmacy. Some studies have shown improvements in symptoms with polypharmacy. One such trial compared clozapine with combination clozapine and risperidone in the management of severe treatment-resistant schizophrenia. Forty patients were studied over a 12-week period and an improvement in negative and positive symptoms of schizophrenia without increased rates of agranulocytosis, weight gain or seizures was seen. The authors concluded that the ‘combination appeared to be safe and well tolerated’ while also ‘improving the overall symptoms – both the positive and negative’. ^{Reference Josiassen, Joseph, Kohegyi, Stokes, Dadvand and Paing11} They recognised that the study was limited by its small size and short duration. A further study in treatment-resistant schizophrenia examined the combination of sulpiride and clozapine and again showed a benefit of dual therapy in a small subgroup of patients. ^{Reference Shiloh, Zemishlany, Aizenberg, Radwan, Schwartz and Dorfman-Etrog12} This study was also of short duration (10 weeks), and although there was no increase in extrapyramidal side-effects in the subgroup receiving dual therapy, problems with hyperprolactinaemia were noted. However, the rates and severity of hyperprolactinaemia were similar to those seen in studies looking at sulpiride alone, and were therefore not thought to be related to the addition of clozapine.

Conversely, the evidence of efficacy of polypharmacy combinations in treatment-resistant schizophrenia is not all favourable. One study published in 2004 compared olanzapine monotherapy to olanzapine combined with sulpiride and found no significant difference in positive or negative symptoms. ^{Reference Kotler, Stous, Reznik, Shwartz, Weizman and Spivak13} Interestingly, the patients receiving dual therapy showed improvement in terms of depressive symptoms as measured by the Hamilton Rating Scale for Depression. Other studies examining the combination of risperidone and clozapine showed no significant improvement in positive or negative symptoms, ^{Reference Honer, Thornton, Chen, Chan, Wong and Bergmann14,Reference Yagcioclu, Akbdede, Turgut, Tumuklu, Yazici and Alptekin15} but did show increased rates of sedation, akathisia, hyperprolactinaemia and elevated fasting blood glucose. ^{Reference Gibson, Patel, Lauriello and Buckley16} A further study looking at combination typical and atypical antipsychotic prescription found that there was little evidence of an improvement in outcome and there was a significant increase in adverse effect burden. ^{Reference Taylor, Mir, Mace and Whiskey17} Another study found that use of more than one antipsychotic was associated with an increased incidence of metabolic syndrome, although the increased incidence could not be solely related to antipsychotic usage and was also linked to clinical and demographic factors. ^{Reference Correll, Frederickson, Kane and Manu18} These studies have all added to the uncertainty regarding the efficacy of antipsychotic polypharmacy, and the BNF states that the ‘prescribing of more that one antipsychotic at the same time is not recommended and may constitute a hazard’. ⁸

A further complication when considering antipsychotic polypharmacy is that the combinations used by clinicians are highly varied. This makes direct comparison between polypharmacy groups and monotherapy groups difficult because often the number of individuals on the same combination regimens are small. The fact that polypharmacy combinations are not consistent means that when considering a ‘polypharmacy’ group the drug regimen is usually not standardised and so direct comparison is fundamentally flawed. Given the large number of antipsychotics that are available for prescription, the possibilities for combination prescriptions are vast and the number of trials that would be required to examine the efficacy of all possible combinations would be innumerable.

One possible route to overcome this problem would be to categorise patients on antipsychotic polypharmacy further – for example into those taking two atypical agents and those taking a typical and an atypical. Interestingly, one such study looked at this and found that 36% of patients were prescribed two atypical antipsychotics, whereas the remainder (64%) received an atypical and a typical antipsychotic. ^{Reference Thompson19} An audit carried out by the Prescribing Observatory for Mental Health found that 74% of people prescribed more than one antipsychotic were prescribed an atypical in combination with a typical drug. ¹⁰ These findings point to the possibility that combination of a typical and an atypical antipsychotic may be the preferred polypharmacy practice.

Many experts have suggested that it would be beneficial to ‘define’ a more manageable number of combination therapies, which would benefit from further investigation, with some suggesting a pharmacological approach to designing combination therapy – for instance adding a stronger dopamine receptor antagonist for those who continue to experience problematic positive symptoms. ^{Reference Sernyak and Rosenheck20} Gibson et al examined possible antipsychotic combinations which could be beneficial based upon pharmacological factors and named four potentially beneficial combinations, namely clozapine plus olanzapine, aripiprazole plus quetiapine, quetiapine plus olanzapine, and aripiprazole plus loxapine. ^{Reference Gibson, Patel, Lauriello and Buckley16} These combinations are thought to have a more potent effect on the D₂ receptor than would be seen with each drug used in isolation, thus in theory improving the symptoms of schizophrenia. Other experts suggest that owing to our limited understanding of both the pathology of schizophrenia and antipsychotic pharmacology it would be beneficial to identify patients who are already receiving combination antipsychotic therapy and are doing well, and use these combinations as a baseline for randomisation.

Many authors have reviewed the literature regarding the practice of polypharmacy and have generally concluded that ‘except in cases where an individual has failed to respond to adequate trial of monotherapies including clozapine, antipsychotic polypharmacy has little support in the medical literature’, ^{Reference Gibson, Patel, Lauriello and Buckley16} and the NICE guidelines reflect this. However, NICE does recognise that there ‘are circumstances where patients and clinicians serendipitously hit upon effective combinations’. ⁷ Therefore, polypharmacy should only be considered after a failed period of monotherapy (this would usually encompass a failed trial of clozapine). Interestingly, one study examined previous clozapine prescription in those on polypharmacy and surprisingly found that only 4% had been given a trial of clozapine before being commenced on polypharmacy. ^{Reference Miller and Craig21} This finding, although not reproduced elsewhere in the literature, suggests that polypharmacy is being considered earlier in a patient's management plan than expected and that it is not being reserved for truly treatment-resistant cases.

Mortality and morbidity

There has long been recognition of the increased mortality associated with schizophrenia, with some studies estimating a relative overall mortality twice that of the general population. ^{Reference Tapp, Wood, Secrest, Erdmann, Cubberley and Kilzieh22} The aetiological basis of this is poorly understood and is not explained by increased suicide rates. Individuals with a diagnosis of schizophrenia are at higher risk of cardiovascular morbidity and mortality than those in the general population. ^{Reference Waddington, Yousseff and Kinsella23} People with schizophrenia also have higher rates of new-onset diabetes compared with a general population control group. ^{Reference Curkendall, Mo, Glasser, Rose Stang and Jones24} Initial studies mainly concentrated on the increased risks of these individuals with regard to physical disease leading to so-called ‘natural deaths’. This was thought to be partly explained by dietary and lifestyle factors –heavy smoking, obesity, excess alcohol intake, high systolic and diastolic blood pressure, low levels of exercise and low levels of education.

Another possible reason behind the higher mortality rates seen in people with schizophrenia is the antipsychotic medications themselves. Concerns regarding antipsychotic use and its association with sudden, unexpected death have long been recognised, with prolongation of the QT_c interval and the risk of torsade de pointes a well-recognised phenomenon. The increased risk of sudden cardiac death has long been associated with typical antipsychotics; however, a recent study has confirmed that this increased risk is also seen with atypical drugs. ^{Reference Ray, Chung, Murray, Hall and Stein25} Interestingly, the risk of death increased with increasing dose of antipsychotic (whether typical or atypical). The study concluded that ‘current users of typical and atypical antipsychotic drugs had a similar dose-related incidence of increased cardiac death’. ^{Reference Ray, Chung, Murray, Hall and Stein25} This finding will lead to further debate regarding the safety of antipsychotics and indeed the practice of high-dose prescribing.

The positive impact of long-term antipsychotic use on survival is relatively neglected. Tiihonen et al found that mortality rates in people with schizophrenia who were not prescribed antipsychotics were ten times higher than in those prescribed antipsychotics. ^{Reference Tiihonen, Walhbeck, Lonnqvist, Klaukka, Ioannidis and Volavkan26} Their study found that patients who were prescribed any antipsychotic drug had decreased mortality compared with the ‘no treatment’ group. Interestingly, no difference in the reduction in mortality demonstrated was seen within the individual drugs prescribed. This study subsequently concluded that one possible route to reducing mortality among young people with schizophrenia would be to target patients who did not renew their prescriptions.

Other studies have considered the complex relationship between antipsychotic use and mortality: Enger et al suggested that ‘cardiovascular risk was inversely associated with intensity of use of antipsychotic drugs’, ^{Reference Enger, Weatherby, Reynolds, Glasser and Walker27} i.e. that the mortality may not be a direct result of the antipsychotic. However, Bralet et al suggested that there was a relationship between mortality and the dose of antipsychotics prescribed, with increased mortality seen in patients prescribed higher doses. ^{Reference Bralet, Yon, Loas and Noisette28} However, this finding was only seen in male patients.

As polypharmacy is becoming more widespread, concerns regarding its safety have been raised. Other concerns, including increased rates of side-effects, pharmacokinetic interactions, reduced adherence to complex medication regimens and increased cost, have also been raised. Difficulties also arise regarding dose adjustments that are required depending on response and side-effects. ^{Reference Chen29} More importantly, the long-term effects of antipsychotic polypharmacy have not yet been fully studied and so this too is an area of growing concern. Waddington et al found that ‘the greater the maximum number of antipsychotics given concurrently, the shorter was patient survival’. ^{Reference Waddington, Yousseff and Kinsella23} A more recent study added support to Waddington's findings by demonstrating ‘a graded relationship between the number of neuroleptics prescribed and mortality of those with schizophrenia’; ^{Reference Joukamaa, Heliövaara, Knekt, Aromaa, Raitasalo and Lehtinen30} in fact, they found that ‘those prescribed three antipsychotics simultaneously were twice as likely to die as those who were prescribed only one’. ^{Reference Joukamaa, Heliövaara, Knekt, Aromaa, Raitasalo and Lehtinen30} This study added that this finding ‘could not be explained by coexistent somatic disease or other risk factors known for premature death’. ^{Reference Joukamaa, Heliövaara, Knekt, Aromaa, Raitasalo and Lehtinen30} These findings have added to the growing physical health concerns regarding antipsychotic polypharmacy.

Management

There are recommendations (both biochemical and electrocardiographic) on monitoring protocols for patients who are prescribed high-dose antipsychotic therapy. ^{Reference Taylor, Paton and Kerwin9} However, despite the common occurrence of antipsychotic polypharmacy, there is no guidance regarding monitoring protocols for these patients. In clinical practice, the guidelines for high-dose prescribing are sometimes extra-polated to cases in which more than one antipsychotic is used or where psychotropic combinations with potential cardiac interactions might occur.

There is a lack of guidance on what to do when faced with patients who are taking more than one antipsychotic, in terms of dose reduction or switching. In clinical practice, this often occurs in patients who have been in continuing care mental health units for long periods. The risk of continuing to expose such individuals to the increased morbidity and mortality of antipsychotic polypharmacy v. the potential deterioration in mental health from reverting to monotherapy or even dose reduction needs to be considered by psychiatrists, and presents a clinical conundrum.

To counteract the increasing trend in polypharmacy prescribing in the face of inconclusive evidence, Suzuki et al examined switching from such antipsychotic use to monotherapy. ^{Reference Suzuki, Uchida, Tanaka, Nomura, Takano and Tanabe31} Individuals were often taking three antipsychotic medications and had been doing so for more than 6 months. Of the 44 patients included in the study, 22 (50%) were successfully converted to monotherapy without detriment to their mental health. A further 10 patients deteriorated following monotherapy, therefore adjuvant antipsychotics were required, albeit at lower doses. The authors concluded that most cases of multiple antipsychotic use were avoidable and recommended that polypharmacy ‘should not be overused and should be the exception, to be used when other therapeutic options have failed’. ^{Reference Suzuki, Uchida, Tanaka, Nomura, Takano and Tanabe31} Suzuki et al also studied patients receiving high-dose antipsychotic polypharmacy without improvement in their symptoms who were switched to second-generation (atypical) antipsychotic monotherapy. ^{Reference Suzuki, Uchida, Watanabe, Yagi and Kashima32} Of the 25 individuals recruited into the study, 23 were successfully switched to monotherapy with no detriment to their mental health at the end-point of the study (12 weeks after completion of their switch) and the vast majority showed an improvement in their Global Assessment of Functioning score.

A third study from Japan examined switching 23 patients with schizophrenia who had been taking high doses of multiple antipsychotic drugs to a single atypical antipsychotic regimen; 18 patients developed 36 episodes of psychotic exacerbations during switching, but nevertheless 20 patients were successfully switched. ^{Reference Kawai, Yamakawa, Matsuzaka, Baba, Tachikawa and Hori33} Another study examined rationalising medication in patients who were on combination oral and depot antipsychotics to a single depot antipsychotic. ^{Reference Soni, Sampath, Shah and Krska34} This study found that switching from polypharmacy to a single, equivalent dose of depot medication alone was associated with ‘unacceptable high rates of relapse’. The reasons for this were unclear; however, the authors suggested that it might be due to the ‘unique pharmacokinetics of these drugs or subtle qualitative differences between them’ and concluded that ‘caution is necessary whenever attempts are made to rationalise polypharmacy in schizophrenic patients’. ^{Reference Soni, Sampath, Shah and Krska34}

These studies have limitations in terms of sample size and study duration and show a mixed picture in the effectiveness of switching from polypharmacy to monotherapy; nevertheless, they are examples of the few articles available in the literature and may provide a potentially useful model for future studies.

Discussion

Antipsychotic polypharmacy is a relatively common and growing practice. The evidence base for its use is strongest in severe treatment-resistant schizophrenia, and although it is applied in other clinical settings the evidence of benefit in these patients is less clear. The possible combinations of antipsychotic drug regimens are innumerable and so a guideline similar to the one seen in the report by Gibson et al regarding the more studied combinations would be beneficial to aid physician prescribing. ^{Reference Gibson, Patel, Lauriello and Buckley16} This would help consolidate our understanding of combination therapy and allow a more pragmatic approach to research. Given the unclear evidence underlying polypharmacy, we suggest a number of factors that may be useful to consider.